Download The integration of T cell migration, differentiation and function

REVIEWS
The integration of T cell migration,
differentiation and function
David Masopust and Jason M. Schenkel
Abstract | T cells function locally. Accordingly, T cells’ recognition of antigen, their subsequent
activation and differentiation, and their role in the processes of infection control, tumour
eradication, autoimmunity, allergy and alloreactivity are intrinsically coupled with migration.
Recent discoveries revise our understanding of the regulation and patterns of T cell trafficking
and reveal limitations in current paradigms. Here, we review classic and emerging concepts,
highlight the challenge of integrating new observations with existing T cell classification
schemes and summarize the heuristic framework provided by viewing T cell differentiation
and function first through the prism of migration.
Department of Microbiology,
Center for Immunology,
University of Minnesota
Medical School, Minneapolis,
Minnesota 55455, USA.
Correspondence to D.M.
e-mail: [email protected]
doi:10.1038/nri3442
Published online 19 April 2013
The fundamental problem that T cells must solve is making contact with rare antigen-bearing cells. This is no easy
task, given that a single T cell occupies only one onehundred-trillionth of the volume of an adult human. This
feat is achieved in a timely manner by targeted migration (homing), a process that is coordinately regulated
with T cell differentiation state. Naive T cells recirculate
through the paracortical regions of secondary lymphoid
organs (SLOs), a strategy that maximizes their opportunity for antigen detection. Activated effector T cells disseminate broadly to exert contact-dependent functions.
After antigen clearance, memory T cells remain distributed throughout the organism, providing an anatomically
pervasive network for immunosurveillance. Memory
T cells themselves are quite heterogeneous, with their
functions being integrated with their location.
New approaches and technologies have precipitated
several recent important discoveries related to T cell
migration. These observations, which include visualization of T cell mobility within different lymphoid and
non-lymphoid compartments, refinements in our understanding of T cell egress, the concept of T cell residency,
new insights into the regulation of homing mol­ecule
expression and further dissection of the relationship
between T cell location and function, are shaping models
of lineage differentiation and the perceived requirements
for effective immunosurveillance. This Review summarizes the recent literature, synthesizes emerging concepts
and highlights future directions for the field.
Immunosurveillance by naive T cells
Entry into secondary lymphoid organs. The adult mouse
contains several billion cells, only ~100 of which
are naive T cells with specificity for any single given
peptide–MHC complex 1,2. However, immunization is
sufficient to initiate recognition and subsequent activation of essentially all naive T cells within 3 days1. This
remarkable feat is mediated by the antigen collecting
capacity of SLOs coupled with the selective migration of
naive T cells to these sites.
Enough lymphocytes leave lymph and enter the subclavian vein to replace all blood-borne lymphocytes 11
times daily3. Half a century ago, Gowans traced the fate of
these cells, discovering that small lymphocytes continuously recirculate among blood, SLOs and lymph4 (FIG. 1a).
Naive blood-borne T cells enter SLOs within 30 minutes5,
where they scan antigen-presenting cells (APCs) for cognate antigen before egressing via the efferent lymphatics.
They then pass through the thoracic duct before returning to the blood ~10–20 h later to begin the cycle anew 6,7.
To overcome shear forces in blood, lymphocytes must
tether; this precipitates their rolling on high endothelial venules (HEVs), a specialized type of post-capillary
vascular endothelium present within lymph node paracortical regions. This action is typically mediated via
short-term interactions between CD62L (L‑selectin) on
T cells and 6‑sulpho sialyl Lewis X oligosaccharides that
decorate various core proteins expressed on HEVs8,9.
Peripheral node addressin (PNAd) is a widely used term
to collectively denote CD62L‑reactive ligands within
HEVs. Selectin-mediated tethering affords the opportunity for chemokine sampling via G‑protein-coupled
chemokine receptors expressed by T cells. CC‑chemokine
ligand 21 (CCL21), which is immobilized on HEVs via
binding to heparin glycosaminoglycans, is recognized by
CC‑chemokine receptor 7 (CCR7) and mediates signalling and subsequent activation of lymphocyte functionassociated antigen 1 (LFA1; also known as integrin-αL)10.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 13 | MAY 2013 | 309
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
a
c
Efferent
lymph
Thoracic duct
Non-lymphoid tissue
CCR7
‘Upstream’
lymph node
T cell
Lymph
node
CCR7
dependent
Efferent
lymphatic
b
Afferent
lymphatic
Afferent
lymphatic
Heart
B cell follicle
HEV
Paracortex
(T cell zone)
CD62L
Rolling
Environment
rich in CCL19
and CCL21
HEV
PNAd
S1P
dependent
LFA1
CCR7
Sinus
CCR7
dependent
Activation
Efferent
lymphatic
Environment
rich in S1P
CCL21
d Time in lymph node
Arrest
S1PR1
↑ sensitivity to
S1P
↓ sensitivity to
CCL19 and
CCL21
ICAM1
Time in lymph
CCR7
↑ sensitivity to
CCL19 and
CCL21
↓ sensitivity to
S1P
Figure 1 | T cell recirculation through lymph nodes. a | T cells enter lymph nodes from blood, then exit lymph nodes via
Nature
Reviews
| Immunology
efferent lymph. Efferent lymph collects in the thoracic duct before returning to blood. b | Many
T cells
enter lymph
nodes
from blood within high endothelial venules (HEVs). These specialized post-capillary venules constitutively express homing
molecules that allow rolling, activation and arrest. c | T cells may also enter the lymph node sinuses via the lymphatics,
either from upstream lymph nodes or from non-lymphoid tissues. Regardless of the mechanism of lymph node entry,
further migration into the T cell zone is CC‑chemokine receptor 7 (CCR7) dependent. d | Sphingosine-1‑phosphate (S1P)
is present in blood and lymph, whereas CC‑chemokine ligand 19 (CCL19) is expressed in lymph node T cell zones.
Exposure to either CCL19 or S1P induces transient CCR7 or S1P receptor 1 (S1PR1) desensitization, respectively. This
process ensures that T cells recirculate among lymph nodes, lymph and blood, rather than dwelling indefinitely in one
compartment. ICAM1, intercellular adhesion molecule 1; PNAd, peripheral node addressin.
Activated LFA1 binds intercellular adhesion molecule 1
(ICAM1), and this precipitates T cell arrest within HEVs,
followed by their transmigration into the T cell zone (also
referred to as the paracortex) of lymph nodes. Thus,
CD62L, CCR7 and LFA1 together mediate the requisite
rolling, activation and arrest processes by which naive
T cells selectively enter lymph nodes (FIG. 1b). Each homing receptor is required, thus allowing selective recruitment of specific cell subsets. For example, neutrophils
express CD62L but lack CCR7, and thus are excluded
from routine traffic through lymph nodes11. The requisite
involvement of three distinct homing receptors allows for
considerable combinatorial complexity, as three different
selectins, 18 distinct chemokine receptors and numerous integrin heterodimers that may contribute to homing
have been identified12. As discussed later in this Review,
effector subsets of T cells use distinct combinations of
homing molecules to specifically migrate where needed.
It should be noted that the requirement for CD62L
and CCR7 for SLO entry is not absolute. α4β7 integrin,
which is expressed at low levels among naive T cells,
can mediate rolling via interaction with mucosal vascular addressin cell adhesion molecule 1 (MADCAM1),
which is expressed on the HEVs of gut-associated SLOs13.
Consequently, genetic defects in CD62L do not impair
T cell homing to Peyer’s patches and only modestly
impair migration to mesenteric lymph nodes. Neither
rolling nor CD62L is required for entering the splenic
white pulp owing to the low shear forces in spleen.
Moreover, a small fraction of memory T cells can use
CXC-chemokine receptor 4 (CXCR4) rather than CCR7
to traffic into lymph nodes, although the significance of
this observation is not yet clear 14–16.
After lymph node entry via HEVs, naive T cells may
enter downstream lymph nodes via the lymphatics.
Many lymph nodes are arrayed in chains. Accordingly,
310 | MAY 2013 | VOLUME 13
www.nature.com/reviews/immunol
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
many lymph nodes will have afferent lymphatic inputs
that effectively represent a continuation of the efferent vessels of more peripheral lymph nodes in addition to those inputs that directly drain non-lymphoid
tissues. A recent study demonstrated that when T cells
are injected directly into the lymphatics afferent to the
popliteal lymph node, they traffic through the subcapsular sinus and accumulate in the medullary sinus and
cords17. Whereas this process is putatively selectin- and
chemokine-independent, subsequent chemotaxis to the
paracortex still requires CCR7. Afferent lymphatic entry
is likely to represent a prominent pathway by which
naive T cells enter downstream lymph nodes (FIG. 1c).
T cell egress from non-lymphoid tissues (discussed
below) highlights another context in which they enter
lymph nodes via the afferent lymphatics, although this
pathway may be limited to antigen-experienced T cells.
Motility and egress. In addition to their functions in
T cell recruitment and antigen collection, lymph nodes
must provide an environment that maximizes cell‑to‑cell
collaboration. After T cells enter lymph nodes, chemokinesis within the T cell zone is CCR7 dependent. In addition to CCL21, CCL19 (an alternative ligand for CCR7)
is expressed in the T cell zone and increases T cell
motility 18. This facilitates scanning; it is estimated that
a single dendritic cell (DC) encounters up to 500 T cells
per hour 19. Competing forces are likely to regulate the
longevity of T cell retention within the lymph node.
Sphingosine-1‑phosphate (S1P) induces chemotaxis of
T cells and is abundantly expressed in blood and lymph,
but it is absent from the T cell zone of lymph nodes and
splenic white pulp. As S1P induces transient desensitization of S1P receptor 1 (S1PR1), recent migrants that
enter SLOs from blood or afferent lymph are relatively
insensitive to S1P signals within efferent lymph and
thus are not initially drawn to exit 20,21. Indeed, treatment
with artificial S1P agonists, such as FTY720, transiently
blocks egress and results in accumulation of T cells
within SLOs22,23. Similarly, CCL19 induces desensitization of CCR7 (REFS 24–26). Consequently, the strength
of CCR7‑dependent retention within the T cell zone will
gradually diminish the longer that resting T cells are positioned within SLOs. Contemporaneously, the sensitivity
to S1P emanating from efferent lymph will gradually be
restored, precipitating T cell egress via the lymphatics
(FIG. 1d). Return of T cells to lymph, or eventually blood,
reverses this process, thus permitting continued recirculation. Although further validation is needed, the CCR7–
S1P axis provides an elegant model of how T cells enter
SLOs and spend an appropriate amount of time surveying DCs in the tissue before leaving to sample different
lymph nodes for cognate antigen22.
Inflammation and T cell activation. HEVs are unusual
among post-capillary venules in that they support high
levels of lymphocyte traffic under uninflamed conditions. However, inflammatory signals such as CCL2
and tumour necrosis factor (TNF), which are expressed
either within inflamed (referred to as ‘reactive’) lymph
nodes or within the tissues that they drain27–29, induce
changes in lymph node architecture and homing requirements. These changes increase the probability of rare
naive T cell interaction with cognate antigen and also
allow recruitment of different T cell subsets and other
leukocyte populations. For instance, local inflammation increases the size of the primary feed arteriole and
microvascular circulation, which serve to hasten the rate
of T cell entry into the affected lymph node30–34. Reactive
lymph nodes induce CCR5 expression on populations
of naive CD8+ T cells, which are subsequently drawn to
local CCL3 and CCL4 gradients emanating from sites
of DC and CD4+ T cell collaboration35,36. Similarly, early
effector CD4+ T cells upregulate CXCR3 and localize to
interfollicular and medullary zones in a CXC-chemokine
ligand 10 (CXCL10)-dependent manner 37. Additionally,
CXCL9 expression increases substantially within the
HEVs of reactive nodes, resulting in the CD62L- and
CCR7‑independent recruitment of CXCR3+ effector
T cells38. Thus, lymph-node-homing rules are not static,
but instead undergo dramatic changes under conditions of inflammation that optimize antigen sampling
specifically at sites of pathogen exposure and also allow
infected lymph nodes to serve as effector sites.
In the rare event that a naive T cell encounters its cognate antigen within the SLO, developmental cues spur a
dramatic series of events. Initial T cell receptor (TCR)
activation elicits a transient upregulation in CCR7 and
CD69 expression. CD69 prevents expression of S1PR1
at the cell surface39. This crucial event effectively increases
the dwell time of recently activated T cells within the
supportive environment of the SLO, which provides
the essential instructional cues of antigen, co‑stimulation
and cytokines to promote productive T cell proliferation40.
After a few days, much of the expanded population leaves
the T cell zone of the lymph node. There is evidence that
T cells compete for antigen, and those that leave the
lymph node first are of the lowest affinity, whereas those
that compete more successfully for antigen stay longer
and continue to proliferate41. This process may ensure
that there is a rapid wave of effectors while also allowing
the cells with optimal specificity to amplify the most.
Migration to non-lymphoid tissues
The imperative for migration. Proliferation signals
that induce clonal expansion are coupled with differentiation and migration cues that coordinate function
with localization. Some T cells may remain within the
SLO; for example, CD4+ T follicular helper (TFH) cells
upregulate CXCR5 and migrate to lymphoid follicles to
provide B lymphocyte help42,43. However, most T cells
exit the SLO to seek infected cells in the organism at
large. Indeed, non-lymphoid tissues represent sites of
infections for most pathogens. Unlike B cells, which
can differentiate into plasma cells and exert effector
functions from distant sites, T cells often need to make
direct contact with antigen-bearing cells to exert their
effector functions and contribute to infection control.
Herein lies a fundamental problem; the human host
comprises 1013 somatic cells that occupy approximately
70 litres in volume. By comparison, a resting lymphocyte has a volume of 290 femto­litres44, occupying only
NATURE REVIEWS | IMMUNOLOGY
VOLUME 13 | MAY 2013 | 311
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
a
b
Sunlight
DC
Skin
Activation
stimulus
c
Dietary
vitamin A
Vitamin D
metabolites
Intestine
Vitamin A
metabolites
Lymph node
T cell
Tissue
factors
↑ CD69
↑ CD103
↓ CD122
↓ CD127
↓ LY6C
Further T cell
differentiation
induced in tissue
↑ CCR10
↑ CCR4
↑ P- and E-
↑ α4β7
↑ CCR9
selectin
Homing potential is programmed by:
• Tissue-specific cues
• TCR stimulation strength
• Inflammation in SLOs
• Inflammation in peripheral tissues
Figure 2 | Location dictates homing and differentiation in two discrete phases. a | The location of T cell priming
influences the induction of peripheral homing molecules to the skin and small intestine. T cell
activation
in skin-draining
Nature
Reviews
| Immunology
secondary lymphoid organs (SLOs) preferentially programs the upregulation of skin homing molecules. Vitamin D
metabolites, catalysed by sunlight, may contribute to this process by inducing CC‑chemokine receptor 10 (CCR10)
expression. b | In the small intestine, dietary vitamin A metabolites are processed by αEβ7+ (also referred to as CD103+)
dendritic cells (DCs) and delivered to T cells during activation in gut-draining SLOs, where they promote α4β7 integrin
and CCR9 expression. c | After migrating to non-lymphoid tissues, a second round of developmental cues is provided
by the tissue microenvironment itself. These factors, which have not been completely identified, modify effector and
memory T cell differentiation in situ. TCR, T cell receptor.
~one one-hundred-trillionth of the organism.
Although T cell proliferation increases the chance that
antigen-specific T cell clones will locate cognate antigen, colocalizing effector T cells with infected host cells
is the key biological challenge; increasing the probability of this interaction probably provided a major selective pressure in evolving the organization and design
principles of the immune system. Multiple levels of
trafficking regulation enable this interaction to occur,
and to occur quickly.
Area code hypothesis. It had long been suspected that
T cells home preferentially to sites of infection45. The
development of the multi-step homing paradigm provided a mechanistic basis for such tissue-selective homing. The combinatorial complexity of unique selectin,
chemokine and integrin combinations that a multi-step
homing model affords precipitated the hypothesis of
organ-specific ‘area codes’, wherein each tissue could
display a unique molecular signature required for entry 46.
In support of this model, different homing molecules
are required for migration to SLOs, gut and skin. On
activation, most effector T cells downregulate CD62L
and CCR7, and upregulate homing molecules that
target them to non-lymphoid tissues rather than to
SLOs. Remarkably, the expression of distinct homing
programmes among effector T cells can be shown to
reflect the site of priming. For example, T cells that
undergo activation in gastrointestinal associated lymphoid tissue (GALT) preferentially upregulate α4β7
integrin and CCR9, which both have important roles in
homing to the small intestinal mucosa47,48. This function
is dependent on GALT stromal cells and the capacity
of intestinal αEβ7+ (also referred to as CD103+) DCs to
convert dietary vitamin A metabolites to retinoic acid,
which in turn induces the expression of gut-homing
molecules on activated lymphocytes49–52. By contrast,
T cells that are activated within the skin-draining
inguinal lymph nodes preferentially upregulate functional E‑selectin and P‑selectin ligands as well as CCR4
and/or CCR10, which are associated with homing to
skin53–57. Interestingly, evidence suggests that vitamin D
metabolites, which are induced by exposure to sunlight
and are overrepresented in skin, induce CCR10 while
suppressing gut homing 58 (FIG. 2a). These profound
discoveries provide a mechanistic explanation for how
the site of T cell priming could program homing to local
non-lymphoid tissues.
312 | MAY 2013 | VOLUME 13
www.nature.com/reviews/immunol
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
Limits of the tissue-specific homing model. The transformative paradigm that T cell priming in specific inductive sites results in a homing programme that is biased to
support trafficking to associated non-lymphoid tissues
has substantial and elegant experimental support and
teleological appeal12,59. However, this paradigm was
never intended to be rigid. Tissue entry requirements
are rarely absolute, regulation of non-lymphoid homing
is multi-factorial and there is abundant evidence that
effector T cells seed uninfected tissues.
Indeed, tissue entry requirements vary by context. For
instance, one study demonstrated that in vitro primed
effector T cell migration to the small intestine was
stringently dependent on CCR9 expression. However,
when T cells were primed in vivo, there was significant
intestinal homing that occurred independently of CCR9
(REF. 60). To add to this complexity, an intraperitoneal
route of immunization exhibited less CCR9‑dependence
than oral immunization for homing to the small intestine60. Thus, in vivo priming is complicated and observations do not always conform to predicted outcomes of
a simplified programmed homing model.
Moreover, there are multiple levels of regulation
for effector T cell accumulation within tissues. For
instance, secondary encounters with antigen after effector cells have already migrated into a non-lymphoid
tissue may affect continued division, survival or retention61,62. Antigen dose and strength of stimulation may
affect the induction of tissue-homing molecules63,64.
And inflamed tissues promote effector T cell recruitment independently of priming site65. Indeed, exogenous
application of CXCR3 ligands, which are broadly recognized by effector T cells, is sufficient to recruit effectors
into the mouse lung or female reproductive tract 66,67.
Additional evidence suggests that early effector T cells
migrate between SLOs before entering tissues, indicating
that they may be subject to homing cues from several
different lymphoid environments68. Interestingly, most
early effector CD8+ T cells isolated from murine spleen
transiently express α4β7 after systemic infections or even
upon local infections outside the gut 68–70. Similar observations were made in human peripheral blood following
subcutaneous vaccination with the yellow fever vaccine
YFV‑17D70. In fact, many studies have demonstrated that
extraintestinal infections or immunization routes result
in the establishment of antigen-specific CD8+ T cells
within the small intestine or otherwise demonstrate
that effector T cells disseminate broadly 68–76.
In summary, the oft-quoted conclusion that infection
of a particular tissue is absolutely required to elicit local
effector T cell migration may not always be true. That
said, the evidence for the induction of tissue-specific
homing programmes within distinct SLO environments
is unequivocal, and there is certainly compelling in vivo
evidence for biased homing in relation to immunization
route. It is likely that the degree to which T cell trafficking is promiscuous as opposed to specifically directed
will vary considerably depending on the context of antigen priming and perhaps even among CD8+ and CD4+
T cells. Promiscuous homing may also vary among tissues, with skin and the central nervous system (CNS)
perhaps being particularly selective. The challenge moving forward will be to define the rules and then apply
this knowledge to optimize vaccination strategies that
achieve the establishment of T cell responses in the
appropriate locations.
Action in non-lymphoid tissues. Effector T cells migrate
to non-lymphoid tissues to seek cognate antigen, and
this requires that they interact with large populations
of host cells. Two-photon microscopy is a tremendously
powerful tool that allows dynamic imaging of T cell
responses in living animals77. Although early two-photon
microscopy studies focused on SLOs, this technology is
increasingly being applied to many non-lymphoid tissues, including pancreas, brain, skin, liver, intestine, lung
and even the beating heart in order to visualize immune
responses in vivo in real time61,78–83. Initial studies have
already revealed tantalizing new insights into effector
T cell behaviour in non-lymphoid tissues.
For example, intravital imaging confirmed that after
migration into the CNS, effector T cells that recognize
cognate antigen undergo division in situ61. Hunter and
colleagues recently demonstrated that unlike naive
T cells in SLOs84, effector T cells do not patrol brain via
Brownian walks in a murine Toxoplasma gondii infection
model85. Rather, they randomly scan in patterns known
as Lévy walks, which cover more territory (FIG. 3a).
This strategy is used by a number of animals, including marine predators, birds and monkeys, to increase
foraging efficiency. The presence of CXCL10 within the
inflamed brain increased the T cell motility rate, but did
not alter this migration pattern.
However, crucial questions remain unanswered. For
instance, how many effector T cells are needed to survey
each and every host cell within a given time frame? How
might these numerical requirements vary between different organs and during different states of inflammation
and infection, and even under different effector T cell
modes of action? In this light, the few measurements
available for effector T cell motility in non-lymphoid
tissues indicate that their movement is not particularly
fast; being measured at ~4–10 μm min–1 (REFS 79,85,86).
One study modelled the cytotoxic T lymphocyte (CTL)
concentration required to eliminate a fast-growing
tumour and arrived at the startling conclusion that >3 × 106
antigen-specific CTLs per gram of tumour were required
to mediate tumour regression87. Recent data suggest
that CD4+ T cells can be remarkably more efficient. In a
Leishmania sp. infection model, Bousso and colleagues
demonstrated that effector CD4+ T cells could signal
infected phagocytes to upregulate inducible nitric oxide
synthase (iNOS) and promote clearance of the infection
as long as the phagocytes were within an 80 μm radius of
an effector CD4+ T cell88. This function was dependent
on interferon‑γ (FIG. 3b). Of course, more pathological
T cell effector functions, such as cytolysis, depend on
the specificity that a requirement for target cell contact
provides. Other strategies, including the adoption of a
dendritic morphology by T cells within skin epidermis,
may increase the number of cell–cell contacts and the
efficiency of immunosurveillance79,89 (FIG. 3c).
NATURE REVIEWS | IMMUNOLOGY
VOLUME 13 | MAY 2013 | 313
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
b
T cell
a
Phagocyte
c
Pathogen
Migration
pathway
Skin
Epithelial cell
Brownian
walk
IFNγ
Epidermis
versus
CD4+
T cell
Lévy walk
Infected
phagocyte
Resting
memory T cell
Figure 3 | Means by which T cells within non-lymphoid tissues increase their efficiency of immunosurveillance and
pathogen control. a | Effector T cells scan infected brain via random Lévy walks (bottom pattern), which cover more
Naturesuch
Reviews
| Immunology
territory than Brownian walks (upper pattern)85. b | CD4+ T cells can extend some effector functions,
as secreted
interferon‑γ (IFNγ, in blue), far beyond the point of contact with an infected phagocyte. This contributes to the
elimination of pathogen in target cells that were not in direct contact with effector CD4+ T cells88. c | Resting memory
T cells have a dendritic morphology in the epidermis and simultaneously contact many epithelial cells79,89.
Memory: homing begets differentiation
Following the resolution of an infection, most effector
T cells die, but a fraction of responding T cells differentiate into long-lived memory T cells. Unlike naive
T cells, which have SLO-restricted homing patterns,
antigen-experienced T cells exhibit a much broader anatomical distribution, being present in both SLOs and nonlymphoid tissues. Memory T cells also exhibit remarkable heterogeneity with respect to phenotype, differentiation and function; importantly, these parameters are
associated with their migration patterns and associated
anatomical locations.
Central and effector memory. Historically, analyses of
cell-mediated immune responses were biased towards
populations of cells that can be easily harvested (that is,
from venous blood) or that are present in high concentrations (that is, in SLOs). Secondary T cell responses
were largely viewed as just a faster and more efficacious
recapitulation of a primary immune response. It was
thought that like primary immune responses, secondary immune responses were initiated in SLOs and
depended on a period of proliferation before migration
of the expanded effector T cell populations to sites of
infection. Crucially, in this model T cells would not be
immediately available to intercept pathogens at nonlymphoid tissue points of entry, where they are typically first encountered. Instead, memory T cells would
be latecomers to non-lymphoid tissue sites of pathogen
re‑exposure.
In 1999, a pivotal study from Lanzavecchia and colleagues highlighted heterogeneity among memory CD4+
and CD8+ T cells isolated from human blood90. They
conceptualized this heterogeneity by delineating memory T cells into two subsets: central and effector memory
T cells (TCM cells and TEM cells, respectively). TCM cells
were defined by the expression of lymph-node-homing
molecules, and together with later studies, the data supported a model in which they exhibited the canonical
memory T cell properties traditionally associated with
anamnestic responses; like naive T cells, they recirculated through SLOs, possessed considerable capacity for
interleukin‑2 (IL‑2) synthesis and proliferation upon
TCR stimulation and, after several days of stimulation,
gave rise to abundant effector T cells that migrated to
sites of infection91,92. TEM cells were defined in blood by
the absence of SLO-homing molecules, and thus putatively represented populations that recirculated through
non-lymphoid tissues. Interestingly, many TEM cells constitutively maintained heightened effector-like functions
(such as cytolytic activity among CD8+ T cells) that
could be rapidly expressed without requiring a period
of re‑differentiation. The great elegance of the TCM/TEM
model was that it married functional specialization with
T cell migration patterns and location.
Innumerable studies have demonstrated that a surprisingly large proportion of both CD4+ and CD8+
memory T cells can be recovered from non-lymphoid
tissues. Congruent with the TCM/TEM model, the pheno­
type and functional properties varied considerably
between SLO and non-lymphoid tissue populations of
memory T cells93,94. To emphasize this crucial theme,
tissue location is often highly predictive of memory
T cell phenotype irrespective of antigen-specificity
and stimulation history 95,96. For example, virusspecific memory TCRαβ+ CD8αβ+ T cells isolated from
the intestinal mucosa constitutively retain granzyme
B expression and cytolytic activity after antigen clearance93,97. Moreover, they express low levels of CD127
(also known as IL‑7Rα) and CD122 (also known as
314 | MAY 2013 | VOLUME 13
www.nature.com/reviews/immunol
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
Immediate interception
Who’s fighting
the infection?
Tissueresident
TEM cell
Migratory
TEM cell
Early response
• In situ proliferation?
• Recruitment?
• Retention of
migratory TEM cell?
Delayed response
Infected
Progeny of tissue
reactivated
TCM cell
Lymph
node
TCM cell
Ignorant
Proliferation
Migration
Figure 4 | Location dictates the rapidity by which memory T cells contribute to protection. Immediate pathogen
interception can only be achieved by those memory T cells present in the tissue at the time of infection: namely resident
Nature Reviews | Immunology
and migrating subsets. If the infection is not immediately controlled, local responses may be rapidly augmented by in situ
proliferation, retention and accumulation of recirculating memory T cells, and recruitment of resting memory T cells from
blood after inflammation. Contemporaneously, central memory T (TCM) cells will be reactivated in draining secondary
lymphoid organs and begin to proliferate and form a new wave of differentiated effectors. After several days, the effector
cell daughters of reactivated TCM cells will migrate to the site of infection in great numbers, making a delayed but robust
contribution to pathogen control. TEM cell, effector memory T cell.
IL‑15Rβ), undergo little homeostatic cell turnover
and exhibit limited proliferation upon reactivation98.
Similar observations have been made for memory
T cells in other non-lymphoid tissues, suggesting that
TEM cells at the frontline are positioned to intercept
pathogens at initial infection sites, where they can provide protection through immediate effector functions.
By contrast, TCM cells, which populate SLOs that drain
frontline tissue sites of infection, are best poised to
regenerate abundant effector responses when needed,
although such a response requires days of proliferation and re‑differentiation. Thus, the broad anatomical distribution of memory T cells and the coupling
of function with location revealed that anamnestic
responses occurred in anatomically and temporally
distinct phases; a paradigm that provides a more
complete explanation of T cell-dependent protective
immunity 91,92,95,96,99 (FIG. 4).
T EM cell recirculation. The interpretation that T EM
cells were in equilibrium between blood and nonlymphoid tissues was consistent with early evidence
suggesting that T cells re‑circulate between the two
compartments. In sheep, memory-phenotype T cells
were shown to be present in afferent lymphatics that
drained peripheral tissues100,101, and in humans some
memory T cells in blood constitutively express skinhoming molecules54; these findings are consistent with
a model of constitutive tissue ingress and egress45. To
test the regulation of egress, T cells were injected into
mouse footpad or lung. Paradoxically, recovery of these
cells from afferent lymphatics was CCR7 dependent 102,103. As CCR7 expression is the cardinal signature
of TCM cells, and TCM cells do not exhibit non-lymphoid
tissue homing characteristics by definition, it is difficult to conceptualize this observation. Nevertheless,
these findings do not appear to be an artefact of the
transfer system, as a large fraction of CD4+ and CD8+
T cells isolated from skin-draining afferent lymphatics
in sheep do indeed express CCR7 (REF. 102). Thus, some
observations are inconsistent with the predictions of
the TCM/TEM model.
TEM cell diversity and limits of the two-subset model.
It has become increasingly clear that a model of a single
recirculating TEM cell subset fails to capture the phenotypical and functional diversity of memory T cells
in non-lymphoid tissues. For instance, CD8+ TEM cells in
the gut express a panoply of markers that are not present
on memory T cells isolated from the blood. This can be
explained by several studies using tissue grafting, intravascular memory T cell transfers and/or parabiosis that
indicate that memory T cells can be long-term residents
in many non-lymphoid tissues, including the small
intestine, skin, brain, lung and salivary glands62,70,104–108.
These data support a model whereby effector T cells
migrate to non-lymphoid tissues, where they then
differentiate in situ while establishing residence without continued recirculation and without representation
in peripheral blood70.
Interestingly, the tissue microenvironment plays a
crucial part in the induction of a tissue-specific pheno­
type post-migration. To illustrate this concept, local
production of transforming growth factor‑β (TGFβ)
induces αEβ7 integrin expression on T cells after their
migration to the small intestinal epithelium (αE is also
NATURE REVIEWS | IMMUNOLOGY
VOLUME 13 | MAY 2013 | 315
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
referred to as CD103). This αEβ7 induction is required
for local T cell maintenance, putatively by binding
E‑cadherin on epithelium97,109. So whereas the SLO
priming environment partially regulates T cell migration to a tissue, once there the tissue provides developmental cues to fine tune compartment-specific T cell
differentiation; a process that putatively optimizes
local adaptive immunity (FIG. 2b). Each tissue is likely to
induce distinct differentiation programmes62,98,105,110,111,
balancing the need for site-specific protective immunity while preserving organ homeostasis. However,
certain signatures do seem common to at least a portion of memory CD8+ T cells established in many tissues, including enhanced expression of CD69, αEβ7 and
granzyme B, and diminished expression of IL‑7Rα and
IL‑15Rβ62,65,97,104–108,110,112,113. Although CD69 expression
occurs transiently on TCR stimulation, this activation
marker is maintained in many non-lymphoid tissues
without the requirement for constitutive cognate antigen recognition65,97,110. CD69 regulates S1P responsiveness, and S1P has been implicated in regulating T cell
egress not only from SLOs but also from non-lymphoid
tissues114. Thus, it is appealing to propose that CD69
could be a marker of tissue residence, although this
remains to be validated more stringently.
The field must now discriminate between recirculating and resident TEM cell subsets. Recent studies
showed that memory CD8+ T cells in the epidermis
were relatively sessile, whereas memory CD4+ T cells
recirculated through the dermis79,115. Previous ovine
studies found that most T cells in the hindlimb draining
afferent lymphatics were CD4+ T cells100,116, suggesting
that CD4+ T cells recirculate more frequently through
peripheral tissues than do CD8+ T cells. However,
CD4+ T cells within the mouse lung can be resident
(and, interestingly, they remain imprinted with lunghoming potential if removed)107. Furthermore, circulating memory CD8+ T cells accumulate in the small
intestinal lamina propria in a β7 integrin-dependent
process117. These data indicate the presence of both
resident and recirculating CD4+ and CD8+ T cells, and
hint that recirculation may be defined by anatomical
compartment (that is, epithelium versus stroma) rather
than by T cell lineage.
Box 1 | Complexity of non-lymphoid tissues
Non-lymphoid tissues are not just ‘bags of cells’ that exist outside lymphoid organs.
Rather, each organ exhibits a unique cell-type composition, organization and function,
and a single organ comprises many distinct tissue types and structures (FIG. 5).
Localization within an organ has an important bearing on T cell recirculation, residence
and location-specific differentiation. Approaches that depend on lymphocyte isolations
and subsequent ex vivo analyses (for example, flow cytometry-based or enzyme-linked
immunosorbent spot (ELISPOT) assays) may contain cells that were present in various
immunologically distinct compartments, including lamina propria, blood, lymph,
tertiary lymphoid aggregates and various types of epithelial surfaces (FIG. 6). To more
fully appreciate T cell immunobiology in non-lymphoid tissues, we will have to consider
not only the organ of origin but also the tissue compartment within an organ. Although
challenging, this goal may be facilitated by the development of new technologies such
as histo-cytometry, a technique that blends the quantitative aspects of flow cytometry
with the localization data from microscopy128.
This raises the importance of appreciating the fine
architecture of tissues. Enzymatic methods for isolating
and studying tissue lymphocytes literally homogenize
what may be several discrete populations that occupy
epithelium and stroma, as well as lymphoid aggregates,
tertiary lymphoid organs, blood and lymphatic vessels
and, in many cases, SLOs (BOX 1; FIGS 5,6). Also, every
tissue is unique. As a cautionary tale, up to 95% (depending on the context) of memory T cells isolated from the
lung were shown to actually be in the microvasculature,
despite perfusion118 (FIG. 6). Restricting analysis to tissue
cells, which are protected from intravascular staining,
will give clearer insight into T cell homing, differentiation and contributions to local protective immunity.
Likewise, most T cells in liver are within sinusoids, which
is contiguous with blood. However, to complicate matters, intravascular sinusoidal and lung microvasculature residence has been documented for natural killer T
(NKT) cells119–121. So, it is possible that blood may be
anatomically heterogeneous and may contain resident
and recirculating subsets.
The original definition of a TEM cell has been complicated by the vast location-dependent heterogeneity and
the distinct migratory patterns of what may be more
precisely described as a collection of several distinct
subsets122. Further refinement is needed to embrace the
more recently discovered complexity.
Anamnestic responses: the primacy of location
T cells require contact with antigen-bearing cells for
triggering effector functions. Infections are typically
initiated at body surfaces, particularly mucosal tissues.
If pathogens are not immediately contained at the site of
entry, pathogen-derived antigens subsequently drain to
SLOs. Accordingly, T cell location dictates how quickly
they may contribute to protective immune responses
(FIG. 4).
For immediate target interception, T cells need to be
pre-positioned at the site of infection. Initial responders
only include resident memory T cells and those migrating T cells that happen to be present within the tissue
at the time of infection. In support of this concept, the
presence of local resident memory T cells in mice greatly
accelerated pathogen clearance in two skin challenge
models104,105. Similar observations have been made in
lung 107. Thus, vaccines will need to establish memory
T cells in tissues if very rapid T cell-mediated control of
the infection is the goal, such as may be the case for HIV,
malaria and tuberculosis vaccines.
Infections induce inflammation, changes in vascular
addressin expression and permeability, and upregulation of chemokine responses in tissues. Although this
is driven by the innate immune system, a recent report
indicates that chemokine induction is potentiated by
in situ reactivation of CD4+ memory T cells123. Topical
application of CXCL9 and CXCL10 to the lumen of
the female reproductive tract in mice recruited other
T cells, although this was limited to effector T cells that
had already been activated in SLOs67. However, there is
evidence that memory T cells can be recruited to sites of
inflammation before antigen-specific reactivation. This
316 | MAY 2013 | VOLUME 13
www.nature.com/reviews/immunol
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
occurs in the lung airways in response to local influenza
infection, in the dorsal root ganglia in response to latent
virus reactivation, and in the skin in response to a local
hypersensitivity reaction79,124,125. Interestingly, if cognate
antigen is recognized in the tissue, recruited memory
T cells may initiate proliferation in situ125. It will be important to define the role that memory (not effector) T cell
migration to sites of inflammation has in accelerating
pathogen control.
If not immediately contained within a non-lymphoid
tissue site of initial exposure, antigen drains via afferent
lymphatics, where it may then be recognized by memory
T cells in downstream SLOs. Such recognition recapitulates the cardinal features of a primary immune response
(albeit more quickly), characterized by T cell proliferation, differentiation and upregulation of peripheral
homing molecules. These processes take several days but
eventually result in a numerically robust tissue-directed
effector immune response (FIG. 4). To conclude, the rapidity by which memory T cells help to control infectious
re‑challenges in tissues is crucially dependent on their
anatomical location92.
Antigen and
pathogen sampling
M cell
Type I
epithelium
Lamina
propria
Type II
epithelium
MALT
Basement
membrane
Afferent
lymph
Capillary
Efferent
lymph
TCM
cell
Resident
TEM cell
Naive
Migratory
TEM cell
Lymph node
Figure 5 | Tissue architecture considerations for T cell trafficking,
subsets
and
Nature Reviews
| Immunology
analyses of isolated populations. An idealized tissue prototype illustrates several
compartments. Epithelium is organized into monolayers (type I) or is stratified (type II).
Mucosal type I epithelial surfaces are often associated with mucosal-associated lymphoid
tissue (MALT). The term MALT is frequently misused129, but only refers to lymphoid
structures that lack afferent lymphatics and sample antigens directly from the lumen of
mucosal tissues via M (microfold) cells. Epithelium is avascular and separated from
connective tissue (referred to as stroma or lamina propria) that contains both blood and
lymphatic vessels. Non-lymphoid tissues drain to lymph nodes via afferent lymphatics.
Each compartment is populated by distinct T cell subsets with distinct functions and
homing patterns. When T cells are isolated from tissue resections or biopsies, T cells are
often pooled from several compartments before analysis, creating an additional source
of heterogeneity. TCM cell, central memory T cell; TEM cell, effector memory T cell.
Closing thoughts
“What kind of music do you usually have here?”
“Oh, we got both kinds. We got country and
western.” The Blues Brothers (1980)
The more I learn, the more I learn how little I know.
Socrates (circa 400 BC)
T cell homing is tightly regulated, and naive, effector
and memory T cells exhibit a variety of migration patterns. Crucially, the regulation of migration is highly
integrated with differentiation and function. However,
this integration is not absolute, and antigen-experienced
T cells exhibit a profound level of heterogeneity that we
have in all likelihood only partly grasped. Challenges
for the field include conceptualizing this diversity while
understanding what it is good for, how it is regulated and
how to manipulate it. Moreover, it is still unclear to what
degree distinct memory T cell populations co­operate
during re‑infection. Is diversity good or will one memory
T cell subset rule them all?
Rational pursuit of these questions will require
accepted and precise definitions. In this, the field has struggled in the past by attributing a broad range of biological
properties (for example, homing, proliferation, cytokine
expression profiles, surface phenotype and lytic activity)
to either TCM or TEM cells122. Different investigators use different criteria. However, these properties are not always
coordinately regulated and cannot be used as synonyms
for the same cell type. For example, expression of CD62L
(the most common marker used to define TCM cells in the
mouse) can be dissociated from proliferation potential126.
Likewise, lack of CD62L expression (one defining signature of TEM cells) does not always predict maintenance
of lytic activity among memory CD8+ T cells (another
defining signature of TEM cells)93. Cells that recirculate
through non-lymphoid tissues express CCR7 upon exit;
are these TCM or TEM cells? This imprecision of definitions
has produced what some have referred to as the Tower
of Babel127. We are not all speaking the same language.
The field is now at a crossroads, as we have newfound
appreciation for additional memory T cell diversity. This
carries the risk of creating additional confusion but also
the opportunity to bring clarity by judiciously applying
appropriate definitions for T cell subsets. As T cell location is antecedent to antigen-recognition, and thus adaptive immune function, the field may be wise to maintain
the spirit of the original TCM/TEM paradigm: viewing T cell
classification through the prism of migration before all
other considerations. In this scheme, TCM cells have a single exact definition: they possess the capacity to migrate
across HEVs under homeostatic (uninflamed) conditions,
whereas TEM cells do not. Importantly, this biological
definition can be inferred on the basis of phenotype.
Unfortunately this nomenclature is not the way the
field is headed, and we are poised to add to the ambiguity. For example, resident memory T cells have now
widely been referred to as ‘TRM cells’ (and we ourselves
have used this term). There are two questions worthy of
reflection before the field settles on this new convention:
NATURE REVIEWS | IMMUNOLOGY
VOLUME 13 | MAY 2013 | 317
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
a
b
c
Vagina
Cervix
iBALT
*
Epithelium
Isolated
lymphoid
follicle
100 µm
100 µm
*
Uterus
Vascular memory
CD8+ T cell
Lamina
propria
100 µm
Ovary
*
*
IEL
100 µm
100 µm
20 µm
100 µm
Figure 6 | Anatomic compartmentalization of representative tissues where T cells may be sampled. a | Lymphocytes
sampled from the intestinal mucosa may include those from epithelium, lamina propria and mucosal-associated
Nature Reviews lymphoid
| Immunology
tissue (MALT). Isolated lymphoid follicle is pictured; >100 isolated lymphoid follicles exist in the mouse large intestine and
terminal ileum. MALT and blood can be contaminating sources of naive T cell and central memory T (TCM) cell preparations from
various non-lymphoid tissues. b | For example, lung may contain inducible broncho-associated lymphoid tissue (iBALT), which
contains high endothelial venules (HEVs) and naive lymphocytes (top panel, red = CD3, blue = B220, green = epithelium). Lung is
also heavily vascularized and T cells are not completely removed by perfusion. The bottom panel illustrates a vascular memory
CD8+ T cell labelled via intravascular staining (cyan) and a CD8+ intraepithelial lymphocyte (IEL) protected from intravascular
staining (blue). Red = epithelium. c | Even a single organ can be quite heterogeneous. For example, in the vagina and ectocervix
the epithelium is stratified squamous, in the uterus it is simple columnar and in the ovary it is cuboidal (‘*’ marks epithelium).
Stromal composition and density also vary among compartments. These complexities are likely to affect differentiation and
maintenance requirements, recirculation patterns and function. White = DAPI and red = collagen IV.
how will we view the past, and how will we conceptualize
the future? If we divorce TRM cells from the TEM cell canon
and anoint them as a completely new subset, will this
bring clarity to the many past studies that described properties of non-recirculating memory T cells but referred
to them as TEM cells? Or will this create additional confusion, particularly for those outside the immediate field of
memory T cell ‘subsetology’? And what will it mean to be
neither a TCM nor a TRM cell? Those cells that remain catalogued under TEM are still quite heterogeneous and poorly
defined transcriptionally. Will we continue to parse them
into completely new ‘lineages’ with entirely new labels?
If so, what will it mean in the future to be a TEM cell?
We would argue that the literature and concepts
will be more accessible if we maintain the TCM and TEM
dichotomy, with each defined precisely as indicated
above. One could claim that central and effector are
inadequate descriptors that carry undesired connotations. Indeed, as the definition of TEM has evolved, it
no longer exclusively refers to memory cells that maintain potent effector functions. Likewise, many TCM cells
rapidly produce effector cytokines and even precipitate
Obar, J. J., Khanna, K. M. & Lefrançois, L.
Endogenous naive CD8+ T cell precursor frequency
regulates primary and memory responses to infection.
Immunity 28, 859–869 (2008).
2.Moon, J. J. et al. Naive CD4+ T cell frequency varies
for different epitopes and predicts repertoire diversity
and response magnitude. Immunity 27, 203–213
(2007).
3. Gowans, J. L. The effect of the continuous re‑infusion of
lymph and lymphocytes on the output of lymphocytes
1.
4.
5.
6.
killing of certain antigen-bearing host cells, and thus
express important effector functions. But it may be far
too late to rebrand these populations with more specific
designations (for example, as THEV+ and THEV– cells) without massive consensus and cooperation among those
that publish in this domain. At this juncture, as new
subsets emerge they could be placed appropriately into
the TCM and TEM taxonomy. Using migration as the primary guiding principle, one might imagine that we will
describe resident TEM cells (those cells that exhibit little
or no recirculation), migrating TEM cells (which routinely
enter and leave tissues) and other subsets (perhaps there
are lymph-node-recirculating populations of TCM cells
that express stem-cell-like qualities, or TEM cell populations that enter neither non-lymphoid tissues nor SLOs
under homeostatic conditions). This nomenclature is
compatible with the literature of the past (which only
bifurcated memory T cells into TCM and TEM subsets) yet
remains flexible enough for the future classification of
emerging or more refined subsets. It has the potential to
be precise, and it defers to the primacy that the location
of a T cell has in the execution of its functions.
from the thoracic duct of unanaesthetized rats.
Br. J. Exp. Pathol. 38, 67–78 (1957).
Gowans, J. L. & Knight, E. J. The route of re‑circulation
of lymphocytes in the rat. Proc. R. Soc. Lond. B. 159,
257–282 (1964).
Whaler, B. C. & Widdicombe, J. G. The blood life-span
of the lymphocyte in rabbits and rats. J. Physiol. 132.
41–42 (1956).
Gowans, J. L. The recirculation of lymphocytes from blood
to lymph in the rat. J. Physiol. 146, 54–69 (1959).
318 | MAY 2013 | VOLUME 13
7.Mandl, J. N. et al. Quantification of lymph node
transit times reveals differences in antigen
surveillance strategies of naïve CD4+ and CD8+ T cells.
Proc. Natl Acad. Sci. USA 38, 263–274 (2012).
8. Gallatin, W. M., Weissman, I. L. & Butcher, E. C.
A cell-surface molecule involved in organ-specific
homing of lymphocytes. Nature 304, 30–34 (1983).
9. Rosen, S. D. Ligands for L‑selectin: homing,
inflammation, and beyond. Annu. Rev. Immunol. 22,
129–156 (2004).
www.nature.com/reviews/immunol
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
10.Bao, X. et al. Endothelial heparan sulfate controls
chemokine presentation in recruitment of lymphocytes
and dendritic cells to lymph nodes. Immunity 33,
817–829 (2010).
11. Andrian, von, U. H. & Mempel, T. R. Homing and
cellular traffic in lymph nodes. Nature Rev. Immunol.
3, 867–878 (2003).
12. Andrian, von, U. H. & Mackay, C. R. T‑cell function and
migration. Two sides of the same coin. N. Engl. J. Med.
343, 1020–1034 (2000).
13. Bargatze, R. F., Jutila, M. A. & Butcher, E. C.
Distinct roles of L‑selectin and integrins α4β7 and
LFA‑1 in lymphocyte homing to Peyer’s patch-HEV
in situ: The multistep model confirmed and refined.
Immunity 3, 99–108 (1995).
14.Scimone, M. L. et al. CXCL12 mediates
CCR7‑independent homing of central memory cells,
but not naive T cells, in peripheral lymph nodes.
J. Exp. Med. 199, 1113–1120 (2004).
15.Okada, T. et al. Chemokine requirements for
B cell entry to lymph nodes and Peyer’s patches.
J. Exp. Med. 196, 65–75 (2002).
16.Hargreaves, D. C. et al. A coordinated change in
chemokine responsiveness guides plasma cell
movements. J. Exp. Med. 194, 45–56 (2001).
17.Braun, A. et al. Afferent lymph-derived T cells and DCs
use different chemokine receptor CCR7‑dependent
routes for entry into the lymph node and intranodal
migration. Nature Immunol. 12, 879–887 (2011).
18. Kaiser, A. A., Donnadieu, E. E., Abastado, J.‑P. J.,
Trautmann, A. A. & Nardin, A. A. CC chemokine ligand
19 secreted by mature dendritic cells increases naive
T cell scanning behavior and their response to rare
cognate antigen. J. Immunol. 175, 2349–2356
(2005).
19. Bousso, P. & Robey, E. Dynamics of CD 8+ T cell
priming by dendritic cells in intact lymph nodes.
Nature Immunol. 4, 579–585 (2003).
20.Lo, C. G. et al. Cyclical modulation of sphingosine1‑phosphate receptor 1 surface expression during
lymphocyte recirculation and relationship to lymphoid
organ transit. J. Exp. Med. 201, 291–301 (2005).
21. Arnon, T. I., Xu, Y., Lo, C., Pham, T. & An, J.
GRK2‑Dependent S1PR1 desensitization is required
for lymphocytes to overcome their attraction to blood.
Science 333, 1898–1903 (2011).
22. Cyster, J. G. & Schwab, S. R. Sphingosine-1‑phosphate
and lymphocyte egress from lymphoid organs.
Annu. Rev. Immunol. 30, 69–94 (2012).
23. Rosen, H., Sanna, M. G., Cahalan, S. M. &
Gonzalez-Cabrera, P. J. Tipping the gatekeeper:
S1P regulation of endothelial barrier function.
Trends Immunol. 28, 102–107 (2007).
24. Bardi, G. G., Lipp, M. M., Baggiolini, M. M. &
Loetscher, P. P. The T cell chemokine receptor CCR7 is
internalized on stimulation with ELC, but not with SLC.
Eur. J. Immunol. 31, 3291–3297 (2001).
25.Kohout, T. A. T. et al. Differential desensitization,
receptor phosphorylation, beta-arrestin recruitment,
and ERK1/2 activation by the two endogenous ligands
for the CC chemokine receptor 7. J. Biol. Chem. 279,
23214–23222 (2004).
26. Zidar, D. A., Violin, J. D., Whalen, E. J. &
Lefkowitz, R. J. Selective engagement of G protein
coupled receptor kinases (GRKs) encodes distinct
functions of biased ligands. Proc. Natl Acad. Sci. USA
106, 9649–9654 (2009).
27.Palframan, R. T. et al. Inflammatory chemokine
transport and presentation in HEV: a remote control
mechanism for monocyte recruitment to lymph nodes
in inflamed tissues. J. Exp. Med. 194, 1361–1373
(2001).
28.McLachlan, J. B. et al. Mast cell-derived tumor
necrosis factor induces hypertrophy of draining
lymph nodes during infection. Nature Immunol.
4, 1199–1205 (2003).
29.Janatpour, M. J. et al. Tumor necrosis factordependent segmental control of MIG expression by
high endothelial venules in inflamed lymph nodes
regulates monocyte recruitment. J. Exp. Med. 194,
1375–1384 (2001).
30.Webster, B. et al. Regulation of lymph node vascular
growth by dendritic cells. J. Exp. Med. 203,
1903–1913 (2006).
31.Soderberg, K. A. et al. Innate control of adaptive
immunity via remodeling of lymph node feed arteriole.
Proc. Natl Acad. Sci. USA 102, 16315–16320
(2012).
32. Hay, J. B. & Hobbs, B. B. The flow of blood to lymph
nodes and its relation to lymphocyte traffic and the
immune response. J. Exp. Med. 145, 31–44 (2012).
33. Steeber, D. A., Erickson, C. M., Hodde, K. C. &
Albrecht, R. M. Vascular changes in popliteal lymph
nodes due to antigen challenge in normal and lethally
irradiated mice. Scann. Microsc. 1, 831–839 (1987).
34. Cahill, R. N., Frost, H. & Trnka, Z. The effects of
antigen on the migration of recirculating lymphocytes
through single lymph nodes. J. Exp. Med. 143,
870–888 (1976).
35.Castellino, F. et al. Chemokines enhance immunity by
guiding naive CD8+ T cells to sites of CD4+ T cell–
dendritic cell interaction. Nature 440, 890–895
(2006).
36.Hickman, H. D. et al. Chemokines control naive
CD8+ T cell selection of optimal lymph node antigen
presenting cells. J. Exp. Med. 208, 2511–2524
(2011).
37.Groom, J. R. et al. CXCR3 chemokine receptor-ligand
interactions in the lymph node optimize CD4+
T helper 1 cell differentiation. Immunity 37,
1091–1103 (2012).
References 35–37 demonstrate the highly
orchestrated localization of T cells to interfollicular
and medullary zones within lymph nodes in
response to antigen challenge.
38.Guarda, G. et al. L‑selectin-negative CCR7− effector
and memory CD8+ T cells enter reactive lymph nodes
and kill dendritic cells. Nature Immunol. 8, 743–752
(2007).
This study demonstrates that migration into
reactive lymph nodes is transiently CD62L and
CCR7 independent.
39.Shiow, L. R. et al. CD69 acts downstream of
interferon-alpha/beta to inhibit S1P1 and lymphocyte
egress from lymphoid organs. Nature 440, 540–544
(2006).
40.Mescher, M. F. et al. Signals required for programming
effector and memory development by CD8+ T cells.
Immunol. Rev. 211, 81–92 (2006).
41. Zehn, D., Lee, S. Y. & Bevan, M. J. Complete but
curtailed T‑cell response to very low-affinity antigen.
Nature 458, 211–214 (2009).
42. Ma, C. S., Deenick, E. K., Batten, M. & Tangye, S. G.
The origins, function, and regulation of T follicular
helper cells. J. Exp. Med. 209, 1241–1253 (2012).
43.León, B. et al. Regulation of TH2 development by
CXCR5+ dendritic cells and lymphotoxin-expressing
B cells. Nature Immunol. 13, 681–690 (2012).
44. Segel, G. B., Cokelet, G. R. & Lichtman, M. A.
The measurement of lymphocyte volume: importance
of reference particle deformability and counting
solution tonicity. Blood 57, 894–899 (1981).
45. Butcher, E. C. & Picker, L. J. Lymphocyte homing
and homeostasis. Science 272, 60–66 (1996).
46. Springer, T. A. Traffic signals for lymphocyte
recirculation and leukocyte emigration: the multistep
paradigm. Cell 76, 301–314 (1994).
47. Salmi, M. & Jalkanen, S. Lymphocyte homing to the
gut: attraction, adhesion, and commitment. Immunol.
Rev. 206, 100–113 (2005).
48. Johansson-Lindbom, B. & Agace, W. W. Generation of
gut-homing T cells and their localization to the small
intestinal mucosa. Immunol. Rev. 215, 226–242
(2007).
49.Iwata, M. et al. Retinoic acid imprints gut-homing
specificity on T cells. Immunity 21, 527–538 (2004).
This paper defined a mechanism for imprinting
T cells with tissue-specific homing.
50.Johansson-Lindbom, B. et al. Functional specialization
of gut CD103+ dendritic cells in the regulation of
tissue-selective T cell homing. J. Exp. Med. 202,
1063–1073 (2005).
51. Stagg, A. J., Kamm, M. A. & Knight, S. C. Intestinal
dendritic cells increase T cell expression of α4β7
integrin. Eur. J. Immunol. 32, 1445–1454 (2002).
52.Hammerschmidt, S. I. et al. Stromal mesenteric lymph
node cells are essential for the generation of guthoming T cells in vivo. J. Exp. Med. 205, 2483–2490
(2008).
53.Austrup, F. et al. P− and E‑selectin mediate
recruitment of T‑helper‑1 but not T‑helper‑2 cells
into inflamed tissues. Nature 385, 81–83 (1997).
54.Picker, L. J. et al. Control of lymphocyte recirculation
in man. II. Differential regulation of the cutaneous
lymphocyte-associated antigen, a tissue-selective
homing receptor for skin-homing T cells. J. Immunol.
150, 1122–1136 (1993).
55. Reiss, Y., Proudfoot, A. E., Power, C. A., Campbell, J. J.
& Butcher, E. C. CC chemokine receptor (CCR)4 and
the CCR10 ligand cutaneous T cell-attracting
chemokine (CTACK) in lymphocyte trafficking to
inflamed skin. J. Exp. Med. 194, 1541–1547 (2001).
NATURE REVIEWS | IMMUNOLOGY
56. Campbell, D. J. & Butcher, E. C. Rapid acquisition of
tissue-specific homing phenotypes by CD4+ T cells
activated in cutaneous or mucosal lymphoid tissues.
J. Exp. Med. 195, 135 (2002).
57. Sigmundsdottir, H. & Butcher, E. C. Environmental
cues, dendritic cells and the programming of tissueselective lymphocyte trafficking. Nature Immunol.
9, 981–987 (2008).
58.Sigmundsdottir, H. et al. DCs metabolize sunlightinduced vitamin D3 to ‘program’ T cell attraction to
the epidermal chemokine CCL27. Nature Immunol.
8, 285–293 (2007).
59. Kunkel, E. J. & Butcher, E. C. Chemokines and the
tissue-specific migration of lymphocytes. Immunity 16,
1–4 (2002).
60. Stenstad, H., Svensson, M., Cucak, H., Kotarsky, K. &
Agace, W. W. Differential homing mechanisms regulate
regionalized effector CD8αβ+ T cell accumulation
within the small intestine. Proc. Natl Acad. Sci. USA
104, 10122–10127 (2007).
61.Kang, S. S. et al. Migration of cytotoxic lymphocytes
in cell cycle permits local MHC I‑dependent control of
division at sites of viral infection. J. Exp. Med. 208,
747–759 (2011).
62. Wakim, L. M., Woodward-Davis, A. & Bevan, M. J.
Memory T cells persisting within the brain after local
infection show functional adaptations to their tissue
of residence. Proc. Natl Acad. Sci. USA 107,
17872–17879 (2010).
63.Svensson, M. et al. Retinoic acid receptor signaling
levels and antigen dose regulate gut homing receptor
expression on CD8+ T cells. Mucosal Immunol. 1,
38–48 (2008).
64.Román, E. et al. CD4 effector T cell subsets in the
response to influenza: heterogeneity, migration, and
function. J. Exp. Med. 196, 957–968 (2002).
65.Mackay, L. K. et al. Long-lived epithelial immunity by
tissue-resident memory T (TRM) cells in the absence of
persisting local antigen presentation. Proc. Natl Acad.
Sci. USA 109, 7037–7042 (2012).
66. Campanella, G. S. V., Medoff, B. D., Manice, L. A.,
Colvin, R. A. & Luster, A. D. Development of a novel
chemokine-mediated in vivo T cell recruitment assay.
J. Immunol. Methods 331, 127–139 (2008).
67. Shin, H. & Iwasaki, A. A vaccine strategy that protects
against genital herpes by establishing local memory
T cells. Nature 491, 463–467 (2012).
68. Liu, L., Fuhlbrigge, R. C., Karibian, K., Tian, T. &
Kupper, T. S. Dynamic programming of CD8+ T cell
trafficking after live viral immunization. Immunity 25,
511–520 (2006).
69.Masopust, D. et al. Activated primary and memory
CD8 T cells migrate to nonlymphoid tissues regardless
of site of activation or tissue of origin. J. Immunol.
172, 4875 (2004).
70.Masopust, D. et al. Dynamic T cell migration program
provides resident memory within intestinal epithelium.
J. Exp. Med. 207, 553–564 (2010).
71.Kaufman, D. R. et al. Trafficking of antigen-specific
CD8+ T lymphocytes to mucosal surfaces following
intramuscular vaccination. J. Immunol. 181,
4188–4198 (2008).
72.Stevceva, L. et al. Both mucosal and systemic routes
of immunization with the live, attenuated NYVAC/
simian immunodeficiency virus SIV(gpe) recombinant
vaccine result in gag-specific CD8+ T‑cell responses
in mucosal tissues of macaques. J. Virol. 76,
11659–11676 (2002).
73.Li, H. et al. Durable mucosal simian
immunodeficiency virus-specific effector memory
T lymphocyte responses elicited by recombinant
adenovirus vectors in rhesus monkeys. J. Virol. 85,
11007–11015 (2011).
74. Belyakov, I. M., Hammond, S. A., Ahlers, J. D.,
Glenn, G. M. & Berzofsky, J. A. Transcutaneous
immunization induces mucosal CTLs and protective
immunity by migration of primed skin dendritic cells.
J. Clin. Invest. 113, 998–1007 (2004).
75.Marshall, D. R. et al. Measuring the diaspora for virusspecific CD8+ T cells. Proc. Natl Acad. Sci. USA 98,
6313–6318 (2001).
76.Agrewala, J. N. et al. Unique ability of activated
CD4+ T cells but not rested effectors to migrate to
non-lymphoid sites in the absence of inflammation.
J. Biol. Chem. 282, 6106–6115 (2007).
77. Coombes, J. L. & Robey, E. A. Dynamic imaging of
host-pathogen interactions in vivo. Nature Rev.
Immunol. 10, 353–364 (2010).
78.Fife, B. T. et al. Interactions between PD‑1 and PD‑L1
promote tolerance by blocking the TCR-induced stop
signal. Nature Immunol. 10, 1185–1192 (2009).
VOLUME 13 | MAY 2013 | 319
© 2013 Macmillan Publishers Limited. All rights reserved
REVIEWS
79.Gebhardt, T. et al. Different patterns of peripheral
migration by memory CD4+ and CD8+ T cells.
Nature 477, 216–219 (2011).
This study identified both recirculating and
resident memory T cell populations within the
skin, and showed that migration behaviour varied
between the dermis and epidermis.
80.Egen, J. G. et al. Intravital imaging reveals limited
antigen presentation and T cell effector function in
mycobacterial granulomas. Immunity 34, 807–819
(2011).
81.McDole, J. R. et al. Goblet cells deliver luminal antigen
to CD103+ dendritic cells in the small intestine.
Nature 483, 345–349 (2012).
82.Looney, M. R. et al. Stabilized imaging of immune
surveillance in the mouse lung. Nature Methods
8, 91–96 (2010).
83.Li, W. et al. Intravital 2‑photon imaging of leukocyte
trafficking in beating heart. J. Clin. Invest. 122,
2499–2508 (2012).
References 82 and 83 apply intravital two‑photon
microscopy to visualize leukocyte motility within
the lung and the heart.
84. Miller, M. J., Wei, S. H., Parker, I. & Cahalan, M. D.
Two-photon imaging of lymphocyte motility and
antigen response in intact lymph node. Science 296,
1869–1873 (2002).
85.Harris, T. H. et al. Generalized Lévy walks and the role
of chemokines in migration of effector CD8+ T cells.
Nature 486, 545–548 (2012).
References 85 and 88 show various strategies
that increase the efficiency of pathogen
immunosurveillance or control by effector
T cells within non-lymphoid tissues.
86. Boissonnas, A., Fetler, L., Zeelenberg, I. S., Hugues, S.
& Amigorena, S. In vivo imaging of cytotoxic T cell
infiltration and elimination of a solid tumor.
J. Exp. Med. 204, 345–356 (2007).
87.Budhu, S. et al. CD8+ T cell concentration
determines their efficiency in killing cognate
antigen-expressing syngeneic mammalian cells
in vitro and in mouse tissues. J. Exp. Med. 207,
223–235 (2010).
88. Eberl, G., Aebischer, T., Späth, G. F. & Bousso, P.
CD4+ T cells rely on a cytokine gradient to control
intracellular pathogens beyond sites of antigen
presentation. Immunity 37, 147–157 (2012).
89.Ariotti, S. et al. Tissue-resident memory CD8+ T cells
continuously patrol skin epithelia to quickly recognize
local antigen. Proc. Natl Acad. Sci. USA 109,
19739–19744 (2012).
90. Sallusto, F., Lenig, D., Forster, R., Lipp, M. &
Lanzavecchia, A. Two subsets of memory T
lymphocytes with distinct homing potentials and
effector functions. Nature 401, 708–712 (1999).
This study was the first to parse memory T cells
into TCM and TEM cells.
91. Sallusto, F., Geginat, J. & Lanzavecchia, A. Central
memory and effector memory T cell subsets: function,
generation, and maintenance. Annu. Rev. Immunol.
22, 745–763 (2004).
92. Masopust, D. & Picker, L. J. Hidden memories:
frontline memory t cells and early pathogen
interception. J. Immunol. 188, 5811–5817 (2012).
93. Masopust, D., Vezys, V., Marzo, A. L. & Lefrançois, L.
Preferential localization of effector memory cells in
nonlymphoid tissue. Science 291, 2413–2417 (2001).
94. Reinhardt, R. L., Khoruts, A., Merica, R., Zell, T. &
Jenkins, M. K. Visualizing the generation of memory
CD4 T cells in the whole body. Nature 410, 101–105
(2001).
95. Lefrançois, L. & Puddington, L. Intestinal and
pulmonary mucosal T cells: local heroes fight to
maintain the status quo. Annu. Rev. Immunol. 24,
681–704 (2006).
96. Woodland, D. L. & Kohlmeier, J. E. Migration,
maintenance and recall of memory T cells in
peripheral tissues. Nature Rev. Immunol. 9, 153–161
(2009).
97.Casey, K. A. et al. Antigen-independent differentiation
and maintenance of effector-like resident memory
T cells in tissues. J. Immunol. 188, 4866–4875 (2012).
98. Masopust, D., Vezys, V., Wherry, E. J., Barber, D. L. &
Ahmed, R. Cutting edge: gut microenvironment
promotes differentiation of a unique memory CD8 T cell
population. J. Immunol. 176, 2079–2083 (2006).
This study showed that the tissue microenvironment
influences T cell differentiation in situ.
99. Gebhardt, T., Mueller, S. N., Heath, W. R. &
Carbone, F. R. Peripheral tissue surveillance and
residency by memory T cells. Trends Immunol. 34,
27–32 (2013).
100.Mackay, C. R., Marston, W. L. & Dudler, L. Naive and
memory T cells show distinct pathways of lymphocyte
recirculation. J. Exp. Med. 171, 801–817 (1990).
101.Abitorabi, M. A. et al. Differential expression of
homing molecules on recirculating lymphocytes from
sheep gut, peripheral, and lung lymph. J. Immunol.
156, 3111–3117 (1996).
102.Debes, G. F. et al. Chemokine receptor CCR7 required
for T lymphocyte exit from peripheral tissues.
Nature Immunol. 6, 889–894 (2005).
103.Bromley, S. K., Thomas, S. Y. & Luster, A. D.
Chemokine receptor CCR7 guides T cell exit from
peripheral tissues and entry into afferent lymphatics.
Nature Immunol. 6, 895–901 (2005).
References 102 and 103 provide evidence that
T cell egress from non-lymphoid tissues is
regulated by CCR7.
104.Gebhardt, T. et al. Memory T cells in nonlymphoid
tissue that provide enhanced local immunity during
infection with herpes simplex virus. Nature Immunol.
10, 524–530 (2009).
105.Jiang, X. et al. Skin infection generates non-migratory
memory CD8+ TRM cells providing global skin
immunity. Nature 483, 227–231 (2012).
106.Hofmann, M. & Pircher, H. E‑cadherin promotes
accumulation of a unique memory CD8 T‑cell
population in murine salivary glands. Proc. Natl Acad.
Sci. USA 108, 16741–16746 (2011).
107.Teijaro, J. R. et al. Cutting edge: tissue-retentive lung
memory CD4 T cells mediate optimal protection to
respiratory virus infection. J. Immunol. 187,
5510–5514 (2011).
References 104, 105 and 107 show that resident
TEM cells positioned within non-lymphoid tissues
accelerate pathogen control upon local re‑infection.
108.Lee, Y. T. et al. Environmental and antigen receptorderived signals support sustained surveillance of the
lungs by pathogen-specific cytotoxic T lymphocytes.
J. Virol. 85, 4085–4094 (2011).
109.El‑Asady, R. et al. TGF-β-dependent CD103 expression
by CD8+ T cells promotes selective destruction of the
host intestinal epithelium during graft-versus-host
disease. J. Exp. Med. 201, 1647–1657 (2005).
110. Kohlmeier, J. E., Miller, S. C. & Woodland, D. L. Cutting
edge: antigen is not required for the activation and
maintenance of virus-specific memory CD8+ T cells in
the lung airways. J. Immunol. 178, 4721 (2007).
111.McCully, M. L. et al. Epidermis instructs skin homing
receptor expression in human T cells. Blood 120,
4591–4598 (2012).
112.Sathaliyawala, T. et al. Distribution and
compartmentalization of human circulating and
tissue-resident memory T cell subsets. Immunity 38,
1–11 (2012).
113.Wakim, L. M. et al. The molecular signature of tissue
resident memory CD8 T cells isolated from the brain.
J. Immunol. 189, 3267–3268 (2012).
114.Ledgerwood, L. G. et al. The sphingosine 1‑phosphate
receptor 1 causes tissue retention by inhibiting the
entry of peripheral tissue T lymphocytes into afferent
lymphatics. Nature Immunol. 9, 42–53 (2008).
115. Bromley, S. K., Yan, S., Tomura, M., Kanagawa, O. &
Luster, A. D. Recirculating memory T cells are a
unique subset of CD4+ T cells with a distinct
phenotype and migratory pattern. J. Immunol. 190,
970–976 (2013).
320 | MAY 2013 | VOLUME 13
116. Abernethy, N. J., Hay, J. B., Kimpton, W. G.,
Washington, E. & Cahill, R. N. Lymphocyte subsetspecific and tissue-specific lymphocyte-endothelial
cell recognition mechanisms independently direct
the recirculation of lymphocytes from blood to lymph
in sheep. Immunology 72, 239–245 (1991).
117.Klonowski, K. D. et al. Dynamics of blood-borne
CD8 memory T cell migration in vivo. Immunity 20,
551–562 (2004).
118.Anderson, K. G. et al. Cutting edge: intravascular
staining redefines lung CD8 T cell responses.
J. Immunol. 189, 2702–2706 (2012).
119.Thomas, S. Y. et al. PLZF induces an intravascular
surveillance program mediated by long-lived LFA‑1–
ICAM‑1 interactions. J. Exp. Med. 208, 1179–1188
(2011).
120.Geissmann, F. et al. Intravascular immune surveillance
by CXCR6+ NKT cells patrolling liver sinusoids.
PLoS Biol. 3, e113 (2005).
121.Scanlon, S. T. et al. Airborne lipid antigens mobilize
resident intravascular NKT cells to induce allergic
airway inflammation. J. Exp. Med. 208, 2113–2124
(2011).
122.Jameson, S. C. & Masopust, D. Diversity in T cell
memory: an embarrassment of riches. Immunity 31,
859–871 (2009).
123.Strutt, T. M. et al. Memory CD4+ T cells induce innate
responses independently of pathogen. Nature Med.
16, 558–564 (2010).
124.Ely, K. H. et al. Nonspecific recruitment of memory
CD8+ T cells to the lung airways during respiratory
virus infections. J. Immunol. 170, 1423–1429
(2003).
125.Wakim, L. M., Gebhardt, T., Heath, W. R. &
Carbone, F. R. Cutting edge: local recall responses
by memory T cells newly recruited to peripheral
nonlymphoid tissues. J. Immunol. 181, 5837–5841
(2008).
126.Hikono, H. et al. Activation phenotype, rather than
central- or effector-memory phenotype, predicts the
recall efficacy of memory CD8+ T cells. J. Exp. Med.
204, 1625–1636 (2007).
127.Rocha, B. & Tanchot, C. The Tower of Babel of CD8+
T‑cell memory: known facts, deserted roads, muddy
waters, and possible dead ends. Immunol. Rev. 211,
182–196 (2006).
128.Gerner, M. Y., Kastenmuller, W., Ifrim, I., Kabat, J. &
Germain, R. N. Histo-cytometry: a method for highly
multiplex quantitative tissue imaging analysis applied
to dendritic cell subset microanatomy in lymph nodes.
Immunity 37, 364–376 (2012).
This report describes an approach for gathering
the cellular positioning information of histological
analyses with the quantitative power of flow
cytometry.
129.Brandtzaeg, P., Kiyono, H., Pabst, R. & Russell, M. W.
Terminology: nomenclature of mucosa-associated
lymphoid tissue. Mucosal Immunol. 1, 31–37
(2008).
Acknowledgements
The authors thank V. Vezys and S. Jameson (University of
Minnesota) for helpful discussion. This study was supported
by U S N a t i o n a l I n s t i t u t e s o f H e a l t h ( N I H ) g ra n t
R01AI084913‑01 (to D.M.), NIH grant T32AI007313 (to
J.M.S.) and the Office Of The Director, NIH, under Award
Number DP2OD006467 (to D.M.). The content is solely the
responsibility of the authors and does not necessarily
represent the official views of the NIH.
Competing interests statement
The authors declare no competing financial interests.
FURTHER INFORMATION
David Masopust’s homepage:
http://masopustvezyslabs.umn.edu/Welcome.html
ALL LINKS ARE ACTIVE IN THE ONLINE PDF
www.nature.com/reviews/immunol
© 2013 Macmillan Publishers Limited. All rights reserved